p-e-w

/

gemma-4-E2B-it-heretic-ara

like 67

ONNX is typically 2.3x faster than the safetensor models. I am wondering if anyone knows if this was ported to ONNX anywhere. Thank you.